Machine learning-based radio frequency (RF) front-end calibration

Abstract

Certain aspects of the present disclosure provide techniques and apparatus for calibrating radio frequency (RF) circuits using machine learning. One example method generally includes calibrating a first subset of RF circuit calibration parameters. Values are predicted for a second subset of RF circuit calibration parameters based on a machine learning model and the first subset of RF circuit calibration parameters. The second subset of RF circuit calibration parameters may be distinct from the first subset of RF circuit calibration parameters. At least the first subset of RF circuit calibration parameters is verified, and after the verifying, at least the first subset of RF circuit calibration parameters are written to a memory associated with the RF circuit.

What is claimed is: 
     
       1. A method for calibrating a radio frequency (RF) circuit, comprising:
 generating, based on a machine learning model and a first subset of RF circuit calibration parameters, values for a second subset of RF circuit calibration parameters; 
 identifying the second subset of RF circuit calibration parameters based on a dropout gradient descent network; and 
 writing at least the first subset of RF circuit calibration parameters to a memory associated with the RF circuit. 
 
     
     
       2. The method of  claim 1 , further comprising verifying at least the first subset of RF circuit calibration parameters, wherein writing the at least the first subset of RF circuit calibration parameters to the memory associated with the RF circuit is performed after the verifying. 
     
     
       3. The method of  claim 1 , wherein the first subset of RF circuit calibration parameters and the second subset of RF circuit calibration parameters are different sets of RF circuit calibration parameters. 
     
     
       4. The method of  claim 1 , wherein identifying the second subset of RF circuit calibration parameters comprises:
 extracting pairwise correlations between parameters in a historical data set of RF circuit calibration parameters; and 
 clustering the pairwise correlations into a plurality of clusters, each cluster of the plurality of clusters being associated with a plurality of parameters in the second subset of RF circuit calibration parameters and a parameter in the first subset of RF circuit calibration parameters. 
 
     
     
       5. The method of  claim 1 , wherein identifying the second subset of RF circuit calibration parameters comprises:
 for each respective pair of parameters including a first parameter and a second parameter in a historical data set of RF circuit calibration parameters: 
 calculating a yield loss generated by calibrating the second parameter using a value of the first parameter; and 
 generating a yield loss similarity value for the respective pair of parameters based on the calculated yield loss for the respective pair of parameters; and 
 clustering pairwise correlations between each respective pair of parameters based on the yield loss similarity value for each respective pair of parameters. 
 
     
     
       6. The method of  claim 1 , wherein identifying the second subset of RF circuit calibration parameters comprises iteratively evaluating RF circuit calibration parameters to identify parameters having a yield loss less than a threshold value. 
     
     
       7. The method of  claim 1 , wherein:
 the dropout gradient descent network comprises a neural network; and 
 identifying the second subset of RF circuit calibration parameters comprises, for each respective parameter of a universe of RF circuit calibration parameters: 
 predicting other parameters in the universe of RF circuit calibration parameters with an identity connection for the respective parameter masked in the neural network; 
 generating a candidate set of predictable parameters based on predicted parameters having a yield loss less than a threshold value; and 
 refining the candidate set of predictable parameters based on a drop probability metric associated with each parameter in the candidate set of predictable parameters. 
 
     
     
       8. The method of  claim 7 , wherein generating the candidate set of predictable parameters comprises:
 assigning a weight for each respective parameter in the candidate set based on weights extracted from the neural network, wherein a weight for the respective parameter corresponds to an effect of the respective parameter on each target parameter; 
 identifying a maximum weight for each respective parameter in the candidate set across a set of target parameters associated with the respective parameter; and 
 initializing a drop probability value for each respective parameter based on a softmax function calculated over the maximum weight for each respective parameter in the candidate set. 
 
     
     
       9. The method of  claim 7 , wherein generating the candidate set of predictable parameters comprises generating the candidate set based on gradient descent optimization of a linear regression function over weights associated with each respective parameter in the candidate set. 
     
     
       10. The method of  claim 1 , wherein the second subset of RF circuit calibration parameters comprises RF circuit calibration parameters predictable with a yield loss less than a threshold value. 
     
     
       11. The method of  claim 1 , further comprising verifying the second subset of RF circuit calibration parameters. 
     
     
       12. The method of  claim 11 , further comprising, after verifying the second subset of RF circuit calibration parameters, writing the second subset of RF circuit calibration parameters to the memory associated with the RF circuit. 
     
     
       13. The method of  claim 1 , further comprising operating the RF circuit based on the second subset of RF circuit calibration parameters and the at least the first subset of RF circuit calibration parameters written to the memory associated with the RF circuit. 
     
     
       14. The method of  claim 1 , wherein the second subset of RF circuit calibration parameters comprises parameters associated with multiple-input, multiple-output (MIMO) techniques. 
     
     
       15. The method of  claim 1 , wherein the second subset of RF circuit calibration parameters comprises parameters associated with higher frequency bands than frequency bands associated with the first subset of RF circuit calibration parameters. 
     
     
       16. The method of  claim 1 , wherein the second subset of RF circuit calibration parameters comprises parameters associated with an increased number of uplinks and downlinks relative to the first subset of RF circuit calibration parameters. 
     
     
       17. The method of  claim 1 , wherein the second subset of RF circuit calibration parameters comprises power usage parameters. 
     
     
       18. The method of  claim 1 , wherein the second subset of RF circuit calibration parameters comprises parameters associated with controlling current biases in the RF circuit. 
     
     
       19. The method of  claim 1 , wherein the second subset of RF circuit calibration parameters comprises parameters for controlling voltages in the RF circuit. 
     
     
       20. A system for calibrating a radio frequency (RF) circuit, comprising:
 memory having executable instructions stored thereon; and 
 one or more processors configured, individually or collectively, to execute the executable instructions in order to cause the system to: 
 generate, based on a machine learning model and a first subset of RF circuit calibration parameters, values for a second subset of RF circuit calibration parameters; 
 identify the second subset of RF circuit calibration parameters based on a dropout gradient descent network; and 
 write at least the first subset of RF circuit calibration parameters to a memory associated with the RF circuit. 
 
     
     
       21. The system of  claim 20 , wherein the one or more processors are further configured, individually or collectively, to cause the system to verify at least the first subset of RF circuit calibration parameters, wherein the one or more processors are configured, individually or collectively, to write the at least the first subset of RF circuit calibration parameters to the memory associated with the RF circuit after the verifying. 
     
     
       22. The system of  claim 20 , wherein the first subset of RF circuit calibration parameters and the second subset of RF circuit calibration parameters are different sets of RF circuit calibration parameters. 
     
     
       23. The system of  claim 20 , wherein the second subset of RF circuit calibration parameters comprises RF circuit calibration parameters predictable with a yield loss less than a threshold value. 
     
     
       24. The system of  claim 20 , wherein the one or more processors are further configured, individually or collectively, to cause the system to verify the second subset of RF circuit calibration parameters. 
     
     
       25. The system of  claim 24 , wherein the one or more processors are further configured, individually or collectively, to cause the system to write, after verifying the second subset of RF circuit calibration parameters, the second subset of RF circuit calibration parameters to the memory associated with the RF circuit. 
     
     
       26. A method for calibrating a radio frequency (RF) circuit, comprising:
 generating, based on a machine learning model and a first subset of RF circuit calibration parameters, values for a second subset of RF circuit calibration parameters; 
 identifying the second subset of RF circuit calibration parameters by iteratively evaluating RF circuit calibration parameters to identify parameters having a yield loss less than a threshold value; and 
 writing at least the first subset of RF circuit calibration parameters to a memory associated with the RF circuit. 
 
     
     
       27. The method of  claim 26 , further comprising verifying at least the first subset of RF circuit calibration parameters, wherein writing the at least the first subset of RF circuit calibration parameters to the memory associated with the RF circuit is performed after the verifying. 
     
     
       28. The method of  claim 26 , wherein the first subset of RF circuit calibration parameters and the second subset of RF circuit calibration parameters are different sets of RF circuit calibration parameters. 
     
     
       29. A system for calibrating a radio frequency (RF) circuit, comprising:
 memory having executable instructions stored thereon; and 
 one or more processors configured, individually or collectively, to execute the executable instructions in order to cause the system to: 
 generate, based on a machine learning model and a first subset of RF circuit calibration parameters, values for a second subset of RF circuit calibration parameters; 
 identify the second subset of RF circuit calibration parameters by iteratively evaluating RF circuit calibration parameters to identify parameters having a yield loss less than a threshold value; and 
 write at least the first subset of RF circuit calibration parameters to a memory associated with the RF circuit. 
 
     
     
       30. The system of  claim 29 , wherein the one or more processors are further configured, individually or collectively, to cause the system to verify at least the first subset of RF circuit calibration parameters, wherein the one or more processors are configured, individually or collectively, to write the at least the first subset of RF circuit calibration parameters to the memory associated with the RF circuit after the verifying.